[email protected] wrote:
These exposure data no shown the illumination question?
It's only giving you an idea of how the camera was metering the scene.
You can have huge differences between models in the way a camera's meter works.
That's one reason you have Exposure Compensation.
If you have a lot of dark areas in an image, many models may try to overexpose it unless you use a -EV setting. If you have a lot of bright areas in an image, many models may try to underexpose it, unless you use a +EV setting.
You have vast differences in metering between camera models. For example, I've noticed that the Sony DSLR-A100 tends to underexpose as much as two stops compared to the Nikon D80 in some conditions (apparently trying to protect the highlights). So, you'd need to use a healthy +EV setting with it to get the same exposure in many conditions.
But, you're only increasing the effective ISO speed if you need to brighten the image. For example, you could shoot at ISO 1600 using a -1EV setting, then brighten the image 1 stop in post processing, giving you the equivalent of ISO 3200 from a shutter speed perspective. That's not what's happening here. The tester probably needed settings equivalent to -1EV in order to get a image exposed correctly (probably due to the darker background).
IOW, you can't draw any conclusions from these images without knowing more about the conditions.
the lighting was the same, then we could start drawing some conclusions. For example, the Canon's ISO 800 setting would appear to be roughly identical to the Pentax's ISO 1600 setting, and the Sony's ISO 400 setting would be almost as sensitive than the Pentax's ISO 1600 setting.
We can tell that by looking at the aperture and shutter speed used and computing the differences in sensitivity for an image exposed for the same brightness level (and to be accurate, you'd need to try and measure patches to tell).
But, we don't know if the lighting was the same or not, and I doubt that the Pentax is a full stop or more worse than the Canon or Sony models for ISO speed sensitivity (even though the aperture and shutter speed settings used to get roughly equivalent brightness would lead you to believe that).
Most DSLR models are going to be within about a half stop of each other in ISO sensitivity for the same setting. It would be nice if they were all calibrated the same way. Some are more or less sensitive than their settings (even though they should all be the same). I think it's unlikely that Pentax "fudged" their ISO speed sensitivity as much as the aperture/shutter speed settings would lead you to believe comparing these models. So, the lighting was probably different for the tests.
IMO, the images don't help much for comparing cameras from an ISO sensitivity/noise perspective. To accurately compare cameras, you need more controlled conditions.